If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Applying a face mask could provoke a trigeminocardiac reflex. We compared the effect of applying bi-nasal prongs with a face mask on breathing and heart rate of preterm infants at birth.
In a retrospective matched-pairs study of infants <32 weeks of gestation, the use of bi-nasal prongs for respiratory support at birth was compared to the use of a face mask. Infants who were initially breathing at birth and subsequently received respiratory support were matched for gestational age (±4 days), birth weight (±300 g), general anaesthesia and gender. Breathing, heart rate and other parameters were collected before and after interface application and in the first 5 min thereafter.
In total, 130 infants were included (n = 65 bi-nasal prongs, n = 65 face mask) with a median (IQR) gestational age of 27+2 (25+3–28+4) vs 26+6 (25+3–28+5) weeks. The proportion of infants who stopped breathing after applying the interface was not different between the groups (bi-nasal prongs 43/65 (66%) vs face mask 46/65 (71%), p = 0.70). Positive pressure ventilation was given more often when bi-nasal prongs were used (55/65 (85%) vs 40/65 (62%), p < 0.001). Heart rate (101 (75–145) vs 110 (68–149) bpm, p = 0.496) and oxygen saturation (59% (48–87) vs 56% (35–84), p = 0.178) were similar in the first 5 min after an interface was applied in the infants who stopped breathing.
Apnoea and bradycardia occurred often after applying either bi-nasal prongs or a face mask on the face for respiratory support in preterm infants at birth.
have shown that applying a face mask can affect breathing in newborns, causing a significant decrease in breathing rate and an increase in tidal volumes or apnoea. A mechanism causing this response could be the trigeminocardiac reflex (TCR), which is an oxygen preserving brainstem reflex, highly prominent in newborns and infants. Stimulation of any branch of the trigeminal nerve along its course can provoke the peripheral TCR and result in apnoea, a sudden decrease in heart rate (HR), changes in blood pressure and gastric hypermotility.
The peripheral TCR can be subdivided into the diving reflex, oculocardiac reflex and the maxillo-mandibular reflex (also referred to as the nasopharyngeal reflex) and can be activated through stimulation of thermic receptors, stretch receptors or nociceptive receptors.
It is possible that the TCR has a stimulation threshold, as a reduction in breathing rate and an increase in tidal volume was not observed when a lightweight cardboard was applied in comparison to a face mask (rim).
Therefore, avoiding the sensitive area around the mouth by using an alternative interface could possibly decrease the chance for inducing the TCR.
We hypothesised that using bi-nasal prongs could reduce the effects of the TCR on breathing and HR in comparison to a face mask. The aim of this retrospective matched-pairs study was to compare the effect of applying bi-nasal prongs with a face mask on breathing and HR of preterm infants.
A retrospective matched-pairs study was performed in infants between 24+0 and 31+6 weeks of gestation. Infants were included from a database containing all recorded resuscitation videos of preterm infants (<32 weeks) receiving respiratory support via bi-nasal prongs or face mask from January 2012 until October 2018 at the General University Hospital in Prague and from March 2014 until October 2018 at Leiden University Medical Centre (LUMC).
Recordings were included if spontaneous breathing was present before the application of an interface. Infants were matched using IBM statistics SPSS version 25 (IBM Software, Chicago, Illinois, USA, 2016) on a 1:1 basis using gestational age (±4 days), birth weight (±300 g), general anaesthesia and gender as the comparators. Recordings were excluded if (1) it was impossible to determine the moment of interface application, (2) the type of interface was changed. A convenience sample was used based on the number of recordings in which bi-nasal prongs were used as respiratory support interface.
In Leiden, respiratory support was provided by the Neopuff™ T-Piece resuscitator (Neopuff™ Infant Resuscitator, Fisher & Paykel Healthcare Ltd., Auckland, New Zealand) via face mask (Neonatal Resuscitation Mask, Fisher & Paykel Healthcare Ltd, Auckland, New Zealand). CPAP was initially given and in case of apnoea and/or bradycardia up to two sustained inflations (SI) for 15 s (PIP 20–25 cm H2O) and positive pressure ventilation (PPV; positive end-expiratory pressure (PEEP) of 5 cm H2O and peak inspiratory pressure (PIP) of 20–25 cm H2O) with a frequency of 40–60/min were given. The fraction of inspired oxygen (FiO2) was initially set at 0.3 and titrated based on the 25th percentile of the Dawson nomogram.
Physiological measurements were collected using a respiratory function monitor (Advanced Life Diagnostics, Weener, Germany) and the Polybench physiological software (Applied Biosignals, Weener, Germany), as described previously.
In Prague, respiratory support was provided by Neopuff™ Infant T-piece Resuscitator (Neopuff™ Infant Resuscitator, Fisher & Paykel Healthcare Ltd., Auckland, New Zealand) via face mask (Laerdal Medical, Stavanger, Norway) or Argyle™ CPAP rubber bi-nasal cannula (Argyle CPAP Nasal Cannula, Covidien, Mansfield, MA, USA). The Argyle™ bi-nasal prongs became part of standard care, as it might avoid frequent complication associated with a face mask (e.g. leak, obstruction and provocation of trigeminal reflexes) and is compatible with the Neopuff. The short and rigid prongs allowed a quick and easy introduction without an increased risk of occlusion due to external pressure of the narrow aperture of the nose. As there was clinical equipoise regarding the two interfaces, the choice for which type of interface to use, was left to the discretion of the caregiver and was made before the infant's condition was evaluated. To minimise leak during nasal support, the nasal cannula leaf's were held in place manually by placing the thumb and the second finger on the upper jaw, while the mouth was closed by a lower jaw thrust using the remaining fingers of the same hand. CPAP was initially given and in case of apnoea and/or bradycardia up to three SI for 15–20 s (1st SI: 20 s, 2nd/3rd SI: 15 s, PIP 20–25 cm H2O) and PPV (PEEP of 5 cm H2O, PIP of 25 cm H2O) with a frequency of 40–60/min were given. The FiO2 was initially set at 0.3 and titrated based on the 25th percentile of the Dawson nomogram.
Physiological measurements were monitored using a Masimo pulse oximeter (Masimo Radical, Masimo Corporation, Irvine, California, USA) and software TRAL (SPM-Service, Zelenograd, Russia). All included infants received respiratory support after cord clamping or cord milking.
Video recordings were reviewed, at normal and half speed, and documented in a case report form.
All the recordings were reviewed by one researcher (KK) and in case of doubt consensus was reached with the help of a second researcher (TL/TM). Gestational age (GA), birth weight, gender, Apgar score at 1 and 5 min after birth, umbilical cord blood pH, mode of delivery, antenatal corticosteroids and the use of general anaesthetics were collected from the medical records of all included infants to describe baseline characteristics.
The primary study outcome was the proportion of infants who were initially breathing at birth and stopped breathing after application of an interface. Initial breathing at birth was defined as chest excursions visible, crying and/or vivid spontaneous movements of the extremities in the time before the interface was applied. Breathing after applying an interface was defined as chest excursions visible 10–60 s after application of an interface or until the start of PPV, when given within 60 s. If infants only took breaths during sustained inflation, they were classified as stopped breathing, since no spontaneous breaths were observed without intervention.
The other study outcomes included the use of SI, PPV, tactile stimulation, suctioning and intubation in the first 5 min after an interface was applied. In addition, HR, oxygen saturation (SpO2) and FiO2 were noted before and 10 s, 30 s, 1 min and 5 min after an interface was applied. SI was defined as a pressure-controlled inflation (20–25 cm H2O) sustained for at least 3 s.
The local institutional Research Ethics Committee of the General University Hospital in Prague and the LUMC approved the study protocol and issued a statement of no objection for performing this study.
Categorical data are presented as n (%), continuous variables are presented as mean ± SD, for normally distributed values, or median (IQR) for non-normal data. Normality was judged from the inspection of histograms. To account for matching, paired tests or mixed effects models were used. The primary analysis was conducted using the related samples McNemar test. Other group comparisons were analysed using the McNemar test (dichotomous variables), the paired samples T-test (normally distributed continuous variables), or the related samples Wilcoxon Rank test (non-normally distributed values). The respective test is mentioned in the corresponding reporting table. A mixed model with a random intercept per individual and matched pair was used to analyse HR, SpO2 and FiO2 over the whole time period. Covariates in the mixed model were group (the variable of interest), time and an indicator variable for time including and beyond 600 s. The indicator variable was added to the model as there was a marked change at that time point for all measurements which was incompatible with a linear trend. To compare HR, SpO2 and FiO2 of the collected time points in the first 5 min after applying the interface measurements are reported per time point as median (IQR). Differences between groups over the whole first 5 min is evaluated using the group covariate of the mixed model (reported as mixed model P-value). Model details are included in a supplement. Mixed models were analysed with R, version 3.5.0 (Foundation for Statistical Computing, Vienna, Austria). All other analyses were conducted in SPSS. A p-value < 0.05 was considered statistically significant.
There were 361 and 70 eligible recordings available with respiratory support via face mask and bi-nasal prongs, respectively. Using the matching criteria, we were able to match 65 infants supported via bi-nasal prongs with 65 infants supported via face mask. All infants in the bi-nasal prongs group were born in Prague, whereas 24/65 (37%) infants in the face mask group were born in Prague and 41/65 (63%) infants were born in Leiden. (Fig. 1) Baseline characteristics were not different between both groups. (Table 1)
Table 1Characteristics of the patients.
Bi-nasal prongs N = 65
Face mask N = 65
Gestational age in weeks, median (IQR)
Birth weight in grams, mean ± SD
927 ± 288
946 ± 268
General anaesthesia, n (%)
Male, n (%)
Complete course of antenatal corticosteroids, n (%)
The proportions of infants who stopped breathing after the interface was applied were not different between the groups (bi-nasal prongs: 43/65 (66%) vs face mask: 46/65 (71%) infants, p = 0.700). SI tended to be given more often (44/65 (68%) vs 33/65 (51%) infants, p < 0.01) and PPV was given more often (55/65 (85%) vs 40/65 (62%) infants, p < 0.001) in the bi-nasal prongs group, without a difference in starting time of PPV after the start of respiratory support (49 (26–82) vs 60 (24–98) seconds, p = 0.446). Tactile stimulation was performed frequently in both groups, while suction was performed more often in the bi-nasal prongs group (53/65 (82%) vs 23/65 (35%) infants, p < 0.001) and was also repeated more often (second time: 23/65 (35%) vs 5/65 (8%) infants, p < 0.001, third time: 6/65 (9%) vs 2/65 (3%) infants, p = 0.289). (Table 2)
There was no statistically significant difference in HR and SpO2 in the first 5 min of the infants who stopped breathing after application of an interface (HR 101 (71–145) vs 110 (68–149) bpm, p = 0.496 (mixed model); SpO2 59% (48–87) vs 56% (35–84), p = 0.178 (mixed model)), but less FiO2 was given in the bi-nasal prongs group at 5 min (0.30 (0.21–0.50) vs 0.56 (0.34–0.98), p < 0.01 (mixed model)) (Fig. 2, Supplementary Table 1).
In this retrospective matched-pairs study we observed that the occurrence of apnoea was high and similar in both groups after applying either the bi-nasal prongs or face mask. This finding suggests that both interfaces, bi-nasal prongs and face mask, are associated with apnoea and/or bradycardia and have the potential of triggering the TCR and thereby increasing the need for PPV. However, to demonstrate whether or not there is a causative relationship mediated through TCR a prospective design would be needed. Nonetheless, we expect the apnoea to be caused by the interface as the apnoea was typically seen right after applying the interface and a previous study showed that infants stopped breathing after a median time of 5 s after applying a face mask.
This is the first study to compare the effect of two interfaces on breathing in preterm infants at birth. We hypothesised that using a different interface which avoided stimulating the sensitive area around the nose and mouth could decrease the effects of the TCR, but were unable to confirm this. The mechanism of the TCR is still relatively unknown in preterm infants and the reason for inducing TCR while using nasal prongs is unclear. Apparently, the threshold to provoke the TCR was still reached or a different pathway was stimulated when nasal prongs were used. It is possible that the bi-nasal prongs provided a sufficient stimulus to the trigeminal nerve by activating the trigeminal receptors inside the nose or the receptors associated with holding the leafs of the bi-nasal prongs to the upper jaw and closing the mouth.
The use of a face mask with a single nasal tube for stabilisation of preterm infants at birth has been previously compared in a randomised controlled clinical trial.
Although, breathing before and after interface application was not investigated, no difference in breathing rate was observed in the first 5 min after the start of respiratory support as well as no differences in SpO2 and HR.
as described in our current study. Nonetheless, trigeminal stimulation and thereby provocation of the TCR might still be reduced when the mouth is gently closed with only one finger and alternative bi-nasal prongs without leaf's and the need for fixation are used.
In this study, more infants stopped breathing after face mask application when compared to our previous study (71% vs 54%).
This is likely the result of the differences in median GA as infants with a lower GA were included in our current study and we previously demonstrated that the occurrence of apnoea is inversely associated with gestation.
observed no difference in PPV in the delivery room when comparing a single nasal tube and a face mask. Although the main cause of these divergent findings is unclear, the decision to start PPV likely varies between caregivers and centres. In this study, all infants in the bi-nasal prongs group were born in Prague, while the majority of infants in the face mask group were born in Leiden. We speculate that the use of respiratory function monitor, which is standard care in Leiden, likely influenced this difference, since it is difficult to identify spontaneous breathing in preterm infants, whereas use of a monitor increases the likelihood of detecting breaths.
To explore the impact of the different centres we compared all infants born in Prague and the occurrence of PPV was not significantly different between groups (bi-nasal prongs vs face mask; 86% (60/70) vs 72% (42/58), p = 0.063). This suggest that the difference in occurrence of PPV might be due to a lower tendency to use PPV in Leiden. It is also noteworthy that some infants stopped breathing after interface placement and started breathing again after SI's, making PPV unnecessary. This explains the discrepancy between the number of infants that stopped breathing after applying a face mask (71%) and the number of infants receiving PPV (62%).
In this study, suctioning was more frequently performed in the bi-nasal prongs group. These results need to be interpreted with the appropriate caution, since this finding might be the result of differences in practice between centres. However, there is a lower threshold for suctioning with bi-nasal prongs, because the respiratory support does not have to be interrupted, in contrast to when a face mask is used.
Although suctioning can lead to vagal-induced bradycardia or apnoea, it is unlikely that this influenced our findings, as we only counted apnoea's that occurred directly after placement of the interface in infants who were breathing and suctioning was usually performed just before the start of PPV or during PPV. However, it is possible that the more frequent use of suctioning could have led to an increased need for PPV in the bi-nasal prongs group as some infants continued breathing after placing the interface, but stopped breathing after suctioning for which then PPV was given.
It is known that applying a face mask (rim) and/or airflow affect breathing in newborns.
This suggests that when CPAP is continuously given directly into the nostrils the cutaneous receptors and the receptors in the nasal cavity of the trigeminal nerve do not reach the threshold for provoking the TCR or the receptors adapt rapidly and the response diminishes with time. Since the primary factor to trigger the TCR is a decrease in intranasal temperature
and CPAP is commonly continuously administered with an air temperature of 37 °C it can be safely used as treatment for AOP. Experimental studies are needed to investigate whether or not starting CPAP could provoke the TCR independently of the interface and if gradually increasing flow or PEEP at the start of CPAP might avoid triggering it.
This was a retrospective matched-pairs study and the results should be interpreted with the appropriate caution. The decision for which interface to use and to start SI, PPV, tactile stimulation, suction and intubation depended on the local protocols and discretion of the caregivers. This also includes the small differences in resuscitation algorithms between the two centres, which might have influenced our results. Analyses of the recordings were dependent on the visibility of the infant's chest and only a small proportion of physiological measurements were available. The infants of whose physiological measurements were available were more likely stable infants as they are more likely to receive respiratory support after the pulse oximetry probe is applied. They also have a better peripheral perfusion, increasing the chance to obtain adequate measurement before the interface is applied. This could have biased our observations and led to an underestimation of the effects on SpO2 and HR.
did show that the majority of the observations on the video could be verified with RFM recordings, which can be considered golden standard. Therefore, we expect the inaccuracy to be minimal. In addition, we used a convenience sample based on availability of recordings where bi-nasal prongs were used to effectuate matching.
In conclusion, there were no differences observed in breathing and heartrate after application of the bi-nasal prongs in comparison to a face mask. Clinicians should be aware that applying an interface could compromise breathing in preterm infants at birth. It is possible that applying an interface stimulates the trigeminal nerve and provoke the TCR as long as the sensitive area around mouth and nose is stimulated. Further studies are warranted to investigate the mechanism of the TCR before an interface can be recommended or developed to minimise the chance for a TCR.
Conflict of interest
None of the authors has financial and personal relationships with other people or organizations that could inappropriately influence (bias) their work. Prof. Dr. A.B. te Pas is recipient of a NWO innovational research incentives scheme (VIDI 91716428).
KK, TL, TM, AP designed the study; KK, TL acquired the data; KK, TM, SB analysed the data; KK, TL, TM, JD, SH, RP and AP were involved with the data interpretation. All authors contributed to the final draft by reviewing the manuscript. No honorarium, grant or other form of payment was given to anyone to produce the manuscript.
Appendix A. Supplementary data
The following are the supplementary data to this article: